Design, synthesis and structural evaluation of peptidomimetics towards foldamers, PAs and non covalent inhibitors of the 20S
proteasome
Dissertation
zur Erlangung des Doktorgrades der Naturwissenschaften Dr. rer. nat.
an der Fakultät für Chemie und Pharmazie
der Universität Regensburg und der Universität Paris XI (Frankreich)
vorgelegt von Andrea Bordessa
aus
Germasino (Italy)
Regensburg 2008
This work was supervised by Prof. Dr. Oliver Reiser in Regensburg and Dr. Sandrine Ongeri and Prof. Dr. Sames Sicsic in Paris
Thesis submission on November 17th, 2008 Thesis defence on December 3rd, 2008
Examination committee: Prof. Dr. Oliver Reiser Dr. Sandrine Ongeri Prof. Dr. Burkhard König Prof. Dr. Sigurd Elz
The following research work was performed from November 2005 to October 2007 in the Institute of Organic Chemistry of the University of Regensburg under the supervision of Prof. Dr.
Oliver Reiser and from November 2007 to October 2008 at the University of Paris XI, in the Laboratoire de Molécules Fluorées et Chimie Médicinale, BioCIS, UMR-CNRS 8076 under the supervision of Dr. Sandrine Ongeri and Prof. Dr. Sames Sicsic.
I would like to thank Prof. O. Reiser, Dr. Sandrine Ongeri and Prof. Sames Sicsic for having given me the opportunity to join their research groups and for their help in these three years.
I also thank the Marie Curie commission for financial support during this PhD programme.
CHAPTER 1 INTROCUTION 1
1.1 Peptide generalities 1
1.2 Primary structure 2
1.3 Secondary structure 2
1.3.1 α-Helix 2
1.3.2 β-sheets 3
1.3.3 turns 4
1.4 Tertiary structure 5
1.5 Quaternary structure 5
1.6 Conformational studies of the secondary structure 6
1.6.1 NMR studies 6
1.6.2 Choice of the solvent 6
1.6.3 2D NMR 6 1.6.4 Hydrogen-deuterium exchange and variation of the temperature 7 1.6.5 Circular dichroism 7 1.6.6 IR in solution 8
1.6.7 X-ray crystallography 9
1.7 Peptidic coupling 9 1.7.1 Coupling reagents 9 1.7.2 Solution phase synthesis 12
1.7.3 Solid phase peptide synthesis (SPPS) 12
1.7.4 Protecting groups 13
1.8 Peptidomimetics 15
CHAPTER 2 δ-AMINO ACIDS TOWARDS FOLDAMERS AND PNAs 17
2.1 Synthesis of cyclic δ-amino acids 17
2.1.1 Three membered rings 17
2.1.4 Four membered rings 20
2.1.5 Five membered rings 22
2.1.6 Six membered rings 27
2.1.7 Bicyclic δ-amino acid 30
2.2 δ-amino acids in foldamers 39
2.3 δ-amino acids in peptide nucleic acids (PNAs) 44
2.3.1 PNA based on aminoethylglicine 44
2.3.2 Linear analogues of Aeg-PNA 48
2.3.3 Cyclic PNAs 52
2.4 Synthesis of δ-amino acids 59
2.4.1 Aim of this work 59
2.4.2 Cyclopropanation 60
2.4.3 Ozonolysis 61
2.4.4 Sakurai allylation 61
2.4.5 Retroaldol lactonisation 62
2.4.6 Introduction of the nitrogen moiety 63
2.4.7 Lactamisation 64
2.4.8 Boc-protection 65
2.4.9 PMB removal by CAN 66
2.4.10 Double bond oxidation 66
2.5 Synthesis of α,δ-pentapeptide 66
2.6 Conformational analysis of the pentapeptide 261 69
2.6.1 IR in solution 69
2.6.2 CD spectroscopy 69
2.6.3 NMR analysis 70
2.6.4 Temperature scan and measurement of the coupling constants 71
2.6.5 2D NMR and molecular modelling studies 72
2.7 Synthesis of α,δ-heptapetide 73
2.7.1 IR in solution 74
2.7.2 CD spectroscopy 75
2.7.3 Temperature scan and measurement of the coupling
constants 76
2.7.4 2D NMR and molecular modelling studies 77
2.8 Synthesis of PNAs 78
2.8.1 Fmoc protection 79
2.8.2 Reduction of the lactone 80
2.8.3 Coupling with thymine 80
2.8.4 PMB removal by CAN 81
2.8.5 Coupling with adenine 82
CHAPTER 3 PROTEASOME AND INHIBITORS 85
3.1 Role of 20S proteasome 85
3.2 Mechanism of the ubiquitin-proteasome pathway 85
3.3 Proteasome inhibitors 86
3.3.1 Covalent inhibitors 86
3.3.2 Peptide aldehydes 91
3.3.3 Peptide boronates 92
3.3.4 Lactacystin and its derivatives 93
3.3.5 Peptide vinyl sulfonates 94
3.3.6 Epoxyketones 95
3.3.7 Non covalent proteasome inhibitors 96
3.4 Biological effect of proteasome inhibitors 100
3.5 Molecular modelling 101
3.5.1 Docking 101
3.5.2 Genetic algorithm 102
3.5.3 Free energy function 104
3.5.4 3D grids 104
3.5.5 Hydrogen bonds 105
3.5.6 The torsional term 105
3.6 Previous works in this lab and aim of this work 106
3.6.1 Literature and crystallographic studies 109
3.6.2 Choice of the docking parameters 111
3.6.3 Docking of the lead molecule and virtual screening of new candidates 116
3.6.4 Synthesis of the fluorinated peptidomimetic 295 120
3.6.5 Solution phase synthesis of the inhibitors 121
3.7 Results and discussion 128
3.8 Conclusions 139
CHAPTER 4 EXPERIMENTAL PART 141
4.1 Instruments and general techniques 141
4.2 Synthesis of the compounds 143
SUMMARY 190
REFERENCES AND NOTES 196
ANNEX I Docking results 208
ANNEX II NMR Data 212
Abbreviations Ac = Acyl
AcOEt = Ethyl acetate Ar = Aryl
Bn = Benzyl
Boc = tert-Butoxycarbonyl Bu = Butyl
CAN = Cerium Ammonium Nitrate Cbz = Benzyloxycarbonyl
CD = Circular Dichroism
COSY =Correlation spectroscopy
DIBAL-H =Diisobutylaluminium Hydride DIPEA = Diisopropylethyl Amine
DMF = Dimethylformamide DMSO = Dimethylsulfoxide DNA = Deoxyribonucleic acid
EDC = Ethyl-N,N-dimethyl-3-aminopropylcarbodiimide ee = Enantiomeric Excess
EI = Electronic impact Eq. = equivalent
Fmoc = 9-Fluorenylmethoxycarbonyl h = hours
HBTU = O-Benzotriazole-N,N,N',N'-tetramethyluronium hexafluorophosphate HOBt = Hydroxybenzotriazol
HOAt = 1-hydroxy-7-azabenzotriazole IR = Infrared spectroscopy
Me = Methyl MeOH = Methanol
m.p. = Melting Point MS = Mass Spectroscopy
NMR = Nuclear Magnetic Resonance NOE = Nuclear Overhauser Effect PNA = Peptide Nucleic Acid PG = Protecting Group Py = Pyridine
PMB = para-Methoxybenzyl
RMSD = Root Mean Square Deviation
ROESY = Rotating Frame NOE Spectroscopy RT = Room Temperature
TFE = Trifluoroethanol
CHAPTER 1 ITRODUCTIO
1.1 Peptide generalities
Peptides (from the Greek πεπτίδια, "small digestibles") are short polymers formed from the linking, in a defined order, of α-amino acids. The link between one amino acid residue and the next is known as an amide bond or a peptide bond. Proteins are polypeptide molecules (or consist of multiple polypeptide subunits). The convention is that peptides are shorter than 50 amino acids residues and polypeptides/proteins are longer. Natural peptides and proteins are mainly composed of 20 α-amino acids to which we can add a few other ones which are relatively rare in nature.
Amino acids are organic molecules which possess an amine and a carboxylic acid. α-amino acids can present different lateral chains leading to molecules with completely different physical properties. The simplest α-amino acid is the glycine, which is the only achiral amino acid.
Natural α-amino acids have been classified in five categories: acidic, neutral, basic, hydrophobic and hydrophilic. When the amino acid is not A glycine, the Cα is a chiral center and natural amino acids are all present in the L-configuration in the nomenclature of Fischer. A few ones extracted from exotic molluscs or in cell walls of some bacterias can be in a D-configuration but their occurrence in nature is anecdotic in comparison to the supremacy of L-amino acids. (Figure 1)
H2N COOH
R COOH
H2N H R
α-amino acid L-amino acid in Fischer representation α
Figure 1
Peptides and proteins present four types of primary structure which will be shortly present above.
1.2 Primary structure
In general, polypeptides are unbranched polymers, so their primary structure can often be specified by the sequence of amino acids along their backbone. However, proteins can become cross-linked, most commonly by disulfide bonds, and the primary structure also requires specifying the cross-linking atoms, e.g., specifying the cysteines involved in the protein's disulfide bonds.
1.3 Secondary structure
The remarkable and highly diverse biological activities exhibited by proteins rely on the unique capacity of these intrinsically flexible chains to fold into well-organized and compact structures.
Linus Pauli, more than half century ago, first understood that detailed information about these molecular and supramolecular structures are a prerequisite for the comprehension of the biological events in the living cell. The formation of tertiary and quaternary structures relies only on a small set of distinct secondary structural elements: sheets, helices and turns (figure 2). Every conformation has is own nomenclature to describe its hydrogen bonding. A non structured conformation is called random coil.
1.3.1 αααα-helix
Helices present a periodic folding having a curly shape. Most of the time, the helix turns clockwise and is called right-handed helix. In the other case, it is a left-handed helix. In the family of helices can be differentiated a few subcategories depending on the periodicity of the helix. The most common is the α-helix, and it is characterised by the presence of an hydrogen
membered ring. Every loop has a length of 0.54 nm and contains 3.6 residues. The dihedral angles ψ and ϕ are between 45 and 60 degrees. The structure is very compact and the lateral chains of the amino acids point out of the helix (figure 1).
Figure 2 α-helix
1.3.2 ββββ-sheet
β-sheet are another typical conformation adopted by proteins. This is a planar conformation
intermolecular hydrogen bonding between the two fragments. The lateral chains of the amino acids point outside the plane, alternatively under and over. It exists 2 different types of β-sheet, the parallel (both fragments are orientated in the same direction) and the anti-parallel, when the fragments are oriented in opposite directions (Figure 3).
parallel ββββ-sheet antiparallel βββ-sheet β Figure 3
1.3.3 turns
Turns are small secondary structures that form an elbow in the peptide sequence and can induce antiparallel β-sheet by placing two fragments in front one to each other. There are classified depending on the ring size of the hydrogen bond forming the turn. β-turns are the most common turns and involve hydrogen bonding between the i residue and the i+3 residue with a ten-
membered ring. There are three types of β-turn depending on the dihedral angles: I, II and III (the type III corresponds to a single turn of 310 helix). The mirror images of these turns are called I’, II’ and III’ (Figure 4).
Figure 4
1.4 Tertiary structure
The tertiary structure of a protein or any other macromolecule is its three-dimensional structure, as defined by the atomic coordinates. So, we can say it is essentially the way as the different secondary structures present in a protein organize them one to each other. For example, some residues which are far in the peptidic chains can be close in the three-dimensional structure.
1.5 Quaternary structure
Many proteins are actually assemblies of more than one polypeptide chain, which in the context of the larger assemblage are known as protein subunits. In addition to the tertiary structure of the subunits, multiple-subunit proteins possess a quaternary structure, which is the arrangement into which the subunits assemble. Enzymes composed of subunits with diverse functions are sometimes called holoenzymes, in which some parts may be known as regulatory subunits and the functional core is known as the catalytic subunit.
1.6 Conformational studies of the secondary structure
It is possible to perform more different analysis for the characterization of the secondary structure.
The most important are the NMR, IR, circular dichroism and X-ray.
1.6.1 MR studies
Since the hydrogen bonds are the principal responsible for the formation of the secondary structure, signals of amide protons are fundamental for the understanding of the peptide organization.
1.6.2 Choice of the solvent
The choice of the solvent is of fundamental importance in the NMR studies. In effect, the solvent can largely affect the ability of a peptide to form a secondary structure, and it is also quite commune that a peptide shows different conformations in different solvents.
1.6.3 2D MR
2D NMR analysis is one of the most powerful tools for the study of the secondary structure of proteins. With unlabelled protein the usual procedure is to record a set of two dimensional homonuclear nuclear magnetic resonance experiments through correlation spectroscopy (COSY), of which several types include conventional correlation spectroscopy and nuclear Overhauser effect spectroscopy (NOESY).1 A two-dimensional nuclear magnetic resonance experiment produces a two-dimensional spectrum. The units of both axes are chemical shifts. The COSY transfers magnetization through the chemical bonds between adjacent protons. The conventional correlation spectroscopy experiment is only able to transfer magnetization between protons on adjacent atoms, so it is transferred among all the protons that are connected by adjacent atoms.
Thus in a conventional correlation spectroscopy, an alpha proton transfers magnetization to the beta protons, the beta protons transfers to the alpha and gamma protons, if any are present, then
experiment is used to build so called spin systems, that is build a list of resonances of the chemical shift of the peptide proton, the alpha protons and all the protons from each residue’s side chain. Which chemical shifts corresponds to which nuclei in the spin system is determined by the conventional correlation spectroscopy connectivities and the fact that different types of protons have characteristic chemical shifts. To connect the different spin systems in a sequential order, the nuclear Overhauser effect spectroscopy experiment has to be used. Because this experiment transfers magnetization through space, it will show crosspeaks for all protons that are close in space regardless of whether they are in the same spin system or not. The neighbouring residues are inherently close in space, so the assignments can be made by the peaks in the NOESY with other spin systems.
1.6.4 Hydrogen deuterium exchange and variation of temperature
This two different studies allows to identify which protons of a molecule are involved in an intramolecular hydrogen bond. In the case of the hydrogen deuterium exchange, to a peptide solved in a solvent which does not present exchangeable deuterium will be added CD3OD. The amide protons in the protein exchange readily with the deuterium of the solvent, so the hydrogen deuterium exchange by NMR spectroscopy follows the disappearance of the amide signals. How rapidly a given amide exchanges reflects its solvent accessibility. Thus amide exchange rates can give information on which parts of the protein are buried, hydrogen bonded etc.
The same principle is at the base of the variation of temperature studies. Generally it is accepted that an amide proton involved in a strong intramolecular hydrogen bond has a low temperature dependence coefficient (∆δ<3 ppb/K), while for a weak intramolecular hydrogen bond is significantly higher (∆δ>8 ppb/K).
1.6.5 Circular dichroism
Circular dichroism (CD) is a form of spectroscopy based on the differential absorption of left- and right-handed circularly polarized light. It can be used to help to determine the structure of macromolecules (including the secondary structure of proteins and the handedness of DNA). CD
solution in high dilution (typically <1 mM, to avoid peptide aggregation) in the absorption band of amide bonds (180-250 nm). Circular dichroism measures the ellipticity of a peptide in this band with UV polarised light. In this band can be observed the absorption π→π* of amides and the absorption will vary according to the hydrogen bonded or non hydrogen bonded state of the amides, therefore it will give information on the presence of a secondary structure. Indeed, a peptide being a chiral molecule, it will present an optical rotation on polarised light but this optical rotation can vary with the wavelength. The analysis is often performed in quartz cells of 1 mm length or less as the solvent absorption can create some parasite noise. All solvents can not be used in CD spectroscopy as the circular dichroism must be measured in a band where the solvent does not absorb, methanol (limit at 195 nm for a cell 1 mm long), trifluoroethanol (TFE) or mixtures methanol/water can be used. Other common organic solvents as THF, acetonitrile, chloroform or dichloromethane can not be used in this case as they absorb in the same region as amides.
The ellipticity θ follows the Beer-Lambert law and can be calculated namely:
θ = CDmeasured/(Cxlxn)
with θ: ellipticity in deg.cm2.dmol-1 C. concentration in mol.L-1
l: length of the cell in dm
n: number of NH in the molecule
This technique has the advantage to be fast and one can see almost immediately if the peptide adopts a secondary structure or not. The limitation is that one can not deduce exactly which amides are involved in hydrogen bonding. It is nevertheless very useful since CD curves of α-peptides are typical of a certain secondary structure and so new peptide curves can be confronted with references. There has been also a lot of work in the field of β-peptides and some references are available but when a peptide containing unnatural amino acids is analysed, this comparison can not be always done with certainty as the curves may differ a lot for a same secondary structure.
1.6.6 IR in solution
Another powerful tools to detect the presence of intramolecular hydrogen is the IR in solution.
The technique consist in the measurement of an IR spectrum at high dilution (usually in dichloromethane or chloroform) to avoid the peptide aggregation. It is so observed the region of amide bond, which present, in the case of intramolecular hydrogen bond, two different bands, one at less than 3400 cm-1 bonded amides and another at more than 3400 cm-1 for non bonded amide.
Limits of this technique are the impossibility to use solvent which can form hydrogen bond with the peptide or which adsorb in the interesting region (DMSO, methanol, etc.).
1.6.7 X-ray crystallography
This is the most potent tools to directly study the secondary conformation of a peptide.
Unfortunately, it is not simple to obtain a crystal of a peptide, especially for the high flexible linear ones. Actually, the crystal structure of more and more peptides of biological interest are available on the Protein Data Bank database.
1.7 Peptidic coupling
Generally speaking, a peptide is formed by the linking of more amino acids by formation of amide bonds. The simple mixture of two or more amino acids in solution at room temperature just brings to the formation of a salt, and the condition to transform this salt into an amide bond it is too harsh for the formation of peptides. Thus, it is necessary therefore to activate the carboxylic group of one of the amino acids so that nucleophilic attack by the amino group of the second can take place, forming the desired amide under mild conditions. In peptide chemistry, this process is called coupling.
1.7.1 Coupling reagents
The most common general coupling method is the use of coupling reagents. This reacts with the free carbonyl group, generating a reactive species, which is not isolated and which is sufficiently
reactive to allow the amide bond formation at room temperature and in mild condition. The most commonly used coupling reagents are the carbodiimides2 (Figure 5).
R N C N R N C N
1 2
Figure 5
Addition of the carboxyl group of the N-protected amino acid to one of the C=N bonds of the carbodiimmide gives the O-acylisourea intermediate 5, the first active species in the coupling reaction. This highly reactive compound can then undergo aminolysis by the amino component, leading to the formation of the amide 6 and the dialkilurea by-product 7.
PGHN O- O
PGHN HN O
R1 C R N
NH+ R
O NH NH
R R
PGHN O
O HN N R
R H2N R1
+
3 4 5
6 7
Figure 6
One of the most important side reaction in the coupling promoted by the carbodiimmides is the formation on an oxazolone intermediate which as to be avoid because it can lead to the racemisation of the substrate (Figure 7).
N O R1
O H R2 R1 N
H
O N
O HR2
O N
N O
R1
O
N O
R1
O
R2 R2
N O
R1
O R2 R1 N
H
NHR O
O R2
8 9 10
12 11
Figure 7
Many of the side reactions that occur when activation is carried out with the carbodiimide alone can be avoid by using some additive. These compounds intercept the O-acylisourea intermediate forming forming a less reactive acylating reagent, which is still potent enough to allow rapid amide bond formation. The most widely used are HOBt (13) and the most reactive, but also expensive HOAt (14).
N N N OH
N N
N N OH
13 14
Figure 8
Other reagent coupling widely used are the uronium salt, in particular HBTU and HATU are currently used. X-ray crystallography demonstrate that both HBTU (15) and HATU (16) crystallized as guanidinium N-oxide (Figure 9).
PF6- X
N N N O
N N
X=CH : 15 X=N : 16
X N
N N+
O
N N
PF6-
Uronium Guanidinium
Figure 9
It is possible to perform a peptidic coupling using two different methods, the solution phase and the solid phase synthesis.
1.7.2 Solution phase peptide synthesis
Despite the larger and larger applications of solid phase methods, the synthesis of peptides in solution remains one of the major chemical approaches to these molecules.3 The principal advantage of this technique is the possibility to isolate and characterize all the intermediates at every step, having a knowledge about the molecular species obtained at every stage. Thus, problems that arise can be immediately identify , when in solid phase synthesis it is possible to do that only after the cleavage to the resin. However, classical peptide synthesis is slower than the solid phase synthesis and it is not suitable for the synthesis of peptide with a large number of residues. In the other hand, it is the most used strategy in some areas, such the high scale synthesis of peptides, the synthesis of peptides composed of unusual or uncommon amino acids (e.g. peptidomimetics) and the synthesis of cyclic peptides.
It exists two different basic strategies: linear or convergent. In the linear strategy the amino acids are coupled singularly to the chain and it is the most suitable method for short peptides. The
of longer peptides. The advantage of this second method is the possibility to prepare in parallel the different scaffold d, but the coupling of the different fragment can be slow and difficult for the steric reasons due to the use of two big fragments in which the reactive group is not so simply accessible.
1.7.3 Solid phase peptide synthesis (SPPS)
Merrifield’s methods for the synthesis of peptides on insoluble polymeric supports has been so succefull that the great majority of peptides are now made using this technique.4-6 The advantages of this technique over classical synthesis in solution phase are those of simplicity and speed of execution. SPPS can be mechanized and has led to the commercialization of automated peptide synthesizers which can be programmed to carry out repetitive steps in the synthesis of a peptide.
In favourable cases, quite complex peptides can be made in a matter of hours by machine-assisted synthesis.
The solid support consists in a polymer chemically inert to all the reagents used in the coupling, insoluble in the reaction’s solvent and simple to handle and to filtrate from liquids. It must be also possible to modify the resin to attach the first amino acid of the synthesis by formation of a covalent bond and remove them at the end of the synthesis. The most common resins are the Merrifield resin, which can be cleaved with HBr or HF and it is compatible with the Boc protecting group, and the Wang resin, which can be cleaved in less drastic conditions (TFA) and it is compatible with the Fmoc protecting group.
1.7.4 Protecting group
In both solution and solid phase synthesis, the choice of the protecting groups plays an important role. It is necessary to choice them to have an orthogonal system, which is a system where it is possible to deprotect easily a group without affecting the others. For the synthesis of peptides the N-protecting group is almost always a urethane derivative. The reason of this choice is the simplicity of the protection and deprotection steps and the possibility, choosing an appropriate urethane (figure 10) to deprotect them under acidic (in the case of Boc group, 17) or basic (in the
19). Usually the Nα is protected by using Boc or Fmoc group (the first is most employed is solution phase synthesis and the second in solid phase), when the more stable Cbz is usually used for the protection of the amino group presents in the lateral chains.
O N
H R O
O N
H R O
O N
H R O
17 18 19
Figure 10
The carboxylic group is usually protected by esterification (Figure 11) forming a methyl or ethyl ester (20, 21 which can be cleaved in basic conditions by aqueous NaOH or LiOH) or an allyl ester (22, cleaved by palladium Tetrakis).
R O
O
R O
O
R O
O
20 21 22
Figure 11
1.8 Peptidomimetics
In recent years, the understanding of the folded and self-assembly processes at work in proteins, which are essentially governed by non-covalent forces, have led to major advances in the de novo design of individual protein secondary structure elements and protein folds from α- polypeptides.7-11 Parallel to this field, chemists have been creating new synthetic oligomers that can self organize spontaneously to form defined secondary structures. These molecules aim at mimicking peptide structure through substances having controlled spatial disposition of functional groups, and for this reason are called peptidomimetics. Peptidomimetics have general features analogous to their parent structure, polypeptides, such as amphiphilicity. They have been developed, to a large extent, for the purpose of replacing peptide substrates of enzymes or peptide ligands of protein receptors.12-15 Peptidomimetic strategies include the modification of amino acid side chains, the introduction of constraints to fix the location of different parts of the molecule,16 the development of templates that induce or stabilize secondary structures of short chains,17, 18the creation of scaffolds that direct side-chain elements to specific locations, and the modification of the peptide backbone. Of these strategies, systematic backbone modifications and structural alterations of the repeat units are most relevant to the field of foldamers. Backbone modifications may involve isosteric or isoelectronic exchange of units or the introduction of additional fragments. For some of these backbones, monomers and sequences giving rise to helical, extended (i.e., “strand”), and turn conformations have been identified. In this field, pseudo-amino acids (β, χ, and δ), due to their ability of these compounds to adopt in solution well-organised secondary structure, can be used as scaffold to place and orient pharmacophores in a predictable manner for the design of molecules with an interesting biological activity. In addition, in contrast with the natural α-peptides, pseudo-peptides display a remarkably in vitro stability to degradation by peptidases from bacterial, fungal and eukaryotic origins (e.g. leucyl aminopeptidase, trypsin, amidase, elastase, 20S proteasome, etc.) which makes them even more attractive for biomedical applications.19 Altogether, such unnatural oligomers designed to reproduce or mimic essential protein structural elements could be of considerable value in the drug discovery.20 In this field, geminal works by the groups of Gellman21 and Seebach19 showed that properties such as folding
peptides containing different amino acids, such as β and χ. Later, it was also demonstrated that the more restricted δ-amino acids can lead to folded oligomers, argument which is exhaustively treated in the next chapter.
CHAPTER 2 δ δ δ δ-AMIO ACIDS TOWARDS FOLDAMERS AD PAs
2.1 Synthesis of cyclic δ δ δ δ-amino acids
In the biomedical research, the synthesis of compound with similar structures to the bioactive peptides (peptidomimetics) is very important to obtain molecules with an improved potency or stability than the natural compound. In this field, cyclic or polycyclic unnatural amino acids, due to the rigidity of the scaffold and to the highly preorganisation of the substituents, are able to offer a conformational bias to obtain a desired structural behaviour as part of oligopeptide or foldamers. Due to the proper spacing between the amino and the carboxylic function δ-cyclic amino acids can be designed as conformationally restricted dipeptide. A first simple classification of cycle δ-amino acids can be done on the base of the number of the atom in the ring. A large part of cyclic amino acids are carbohydrate derivatives bearing both an amine and a carboxylic acid functionality also referred to as sugar amino acid (SAA). In particular furanoid and pyranoid amino acids found a large application due to the cheap sugar starting materials, the simplicity of the synthesis and the possibility of an high functionalisation by reaction of the hydroxyl functions present in the molecule.
2.1.1 Three membered ring δδδδ-amino acids
In 1990 Kaltenbronn et al.22 synthesised a dipeptide isostere replacing the amide bond with an epoxide ring. The key step of the synthesis was the epoxidation of an alkene by mCPBA, obtaining the desired product as a mixture of diastereomers (Figure 12).
BocHN CHO BocHN
SiMe3
BocHN COOH
BocHN O COOH
Ph3CH2CCSiMe3+Br-, n-BuLi
HB , H2O2 1.
2. CDI, MeOH
mCPBA
24 25
27 26
Figure 12: synthesis of epoxy amino acids by Kaltenbronn et al.
After this first example, the same strategy was followed by others groups23-25 obtaining δ-amino acids with different substituents and diastereomeric ratio of the epoxide ring. In 1996 Mann et al.26 (Figure 13) published another synthesis of this promising type of δ-amino acids using a Mukaiyama aldol type reaction between the nonstereogenic sylilketene acetal 31 and the chiral aldehyde 30. The reaction performed in the presence of boron trifluoride etherate gave product 32 in a single diastereomers which was then treated with mCPBA obtaining the epoxide δ-amino ester 33.
H NHCbz
O
MeOCH2PPh3Cl,
NaHMDS, THF NHCbz
OMe
PheSeCl, DCM, -78 °C
CHO NHCbz
SePh OSiMe3
OMe
BF3 Et2O, DCM, -78 °C CbzHN
PhSe
COOMe OH
m-CPBA, K2CO3, MeOH, -15 °C
CbzHN
COOMe H
H
O NaOH
CbzHN
COOH H
H O 28
29
32 30
33 34
31
Figure 13: route for epoxy amino acid by Mann et al.
A different approach to 3 membered ring δ-amino acid has been performed by Wipf in 200527 which synthesised substituted cyclopropane dipeptide isostere (Figure 14). Methyl alkyne 36 was hydrozirconated with Cp2ZrHCl, transmetalated to Me2Zn and added to (diphenylposphinylimino)phenylarene to provide the corresponding allylic amide, which was converted to the desired cyclopropane 37 after treatment with CH2I2. Simultaneous N and O deprotection followed by selective N-Cbz protection afforded the alcohol 38a and 38b as a separable mixture of diastereomers. The desired δ-amino acid 39 was then obtained by a two step oxidation of the hydroxyl function to carboxylic acid.
H OTBDPS
Br Br 1. n-BuLi
2. MeI
OTBDPS Me
OTBDPS Ph2(O)PHN
Ph
1. Cp2ZrHCl 2. Me2Zn 3. PhCHNP(O)Ph2 4. CH2I2 60%
1. HCl 2. CbzCl
OH CbzHN
Ph
OH CbzHN
Ph O
1. Dess-Martin 2. NaClO2, NaH2PO4
35 36
37
38a
39 OH CbzHN
Ph
38b +
Figure 14: synthesis of cyclopropane dipeptide isostere by Wipf et al.
2.1.2 Four membered ring δδδδ-amino acids
Only few works have been published about the synthesis of δ-amino acid with 4 ring atoms. Very active in this field, Fleet’s group published different works reporting the synthesis and/or the secondary structural investigation of δ-2,4 oxetane amino acids.
O HO
OH OH OH
L-arabinose
O
O O OTBDPS
HO
O
O O BnO
O
BnO OTf O O
BnO COOMe
N3
O BnO COOiPr
NH2
1. TDPBSCl, imidazole, DMF 2. CuSO4, acetone
N3
N3 1. NaH, BnBr, Bu4Ni 2. Bu4NF, THF 3. TsCl, pyridine 4. NaN3, DMF
1. TFA/H2O 3:2 2. Br2, BaCO3, dioxane:H2O 3. Tf2O, pyridine, DCM
MeOH, K2CO3
1. iPr, K2CO3 2. H2, Pd/C, AcOEt
40 41 42
44 43 45
Figure 15: route for oxetane amino acid by Fleet et al.
For example in 200828, 29 (Figure 15) 2,4-cis oxetane monomers have been synthesised starting from inexpensive L-arabinose. Key step of the synthetic route is the ring closure of an α-triflate of a γ-lactone in basic methanol, which provide the desired oxetane 44.
O RO COOiPr
H NH2
O RO COOiPr
H NH2
R = H or protective group
46 47
Figure 16: other oxetane amino acids by Fleet. et al.
With the same methodology, but starting from different sugar like L-rhamnose or D-xylose, was possible obtain oxetane with different substituent30, 31 (figure 16).
2.1.3 Five membered ring δδδ-amino acids δ
Figure 17: furanoid amino acid by Smith et al.
In 1999 Smith et al.32 proposed a synthesis of a tetrapeptide based on the trans-5-aminomethyl- tetrahydrofuran-2-carboxylate (figure 17). Treatment of the open chain triflate 49 with methanolic hydrogen chloride give the furan ring via an SN2-like closure of the C-5 hydroxyl onto C-2 with inversion of configuration. Treatment of primary mesylate 50 with sodium azide afforded the azido ester which was reduced by catalytic hydrogenation to give the free amino group.
O O
HO OH
OH
CH2OH
MeOOC O
O
O O
TfO 1. 1% HCl, MeOH 2. MsCl, py, DMAP
O HO
MeOOC
OH
O HO PrOOC
OH O
HO PrOOC
OH
1. NaN3, DMF 2. iPrOH, K2CO3
H2, Pd/C, iPrOH
48 49 50
52 51
OMs
N3 NH2
BnO O BnO
N3
BnO OMe
BnO O BnO
N3
BnO OH
N3
OH OH
OBn OBn
OBn
OH OBn
OBn
OBn O BocN
BnO
COOH OBn BocHN
NaBH4, MeOH
1. PPH3, tol.
2. Boc2O,K2CO3
PDC, DMF HCl
53 54 55
57 56
Figure 18: route for furanoid amino acids by chakraborty et al.
In 2000 Chakraborty et al.33 proposed a synthesis of the furanoid δ-amino acid 57 from the hexose substrate 53, involving an unusual 5-exo opening of the terminal aziridine ring of 56 by the γ-benzyloxy during the oxidation of the primary hydroxyl group by pyridinium dichromate, with a complete stereocontrol of the ring opening under these conditions (Figure 18).
O OAc
BzO OBz BO
O
BzO OBz BzO
O
HO OH
HO
OHC O
O O
MsO O MsO
HO OH
HO O
O OH
HO O
O OH
O
N3
N3
N3 N3
NHBoc
1. Co2(CO)8, DEMS, COgas, DCM 2. AcOH, H2O, THF 3. MsCl, py
1. NaN3, DMF 2.KOtBu, MeOH
1. 2-2-dimethoxypropane, pTsOH, acetone
2. Dess-Martin periodane, DCM
1. CH2O, 2M NaOH 2. NaBH4
3. MsCl, py 1. 4M HCl, THF
2. 2M NaOH, diox
1. Boc-ON, PPh3, THF
2. TEMPO, KBr, NaOCl, NaHCO3
58 59 60
61 63 62
64
OMs
An original structure was synthesised by Van Well et al.34 in 2003, which synthesised a locked furanoid amino acid (figure 19): the key step of the synthesis is the regioselective ring closure of the dimesylate 62 to exclusively afford the four-membered ring.
OH NH2 COOH
O
O O
O O
OTBDPS
O OTBDPS
O OHC O
O2N H OH
O
OTBDPS H2N
OTBS H
O COOMe H2N
OTBS H 1. NaNO2, H2SO4, H2O
2. BH3-Me2S, THF
3. TBDPSiCl, Et3N, DMAP, DCM 4. LiHMDS, PhSeBr, THF
OTBDPS
CuI, CH2CHMgBr, Me2S, THF
1. DIBAL-H, DCM 2. Et3SiH, BF3-Et2O
O3, PPH3, DCM 3-methyl-1-nitrobutane,
LaLi3(R-binol)3-LiOH, THF
OTBDPS OTBDPS
1. TBSOTf, 2-6-loutidine, DCM 2. H2, Ni Raney, H2PtCl6, MeOH
1. Boc2O, Et3N, THF 2. PDC, DMF 3. CH2N2, DCM
65 66 67
68 70 69
71 72
Figure 20: route for furanoid amino acid by Hanessian et al.
In 2004 Hanessian and Brassard35 (figure 20) synthesised a constrained oxacyclic hydroxyethilene isostere of aspartyl protease inhibitors. Introduction of the nitrogen moiety was performed by a nitroaldol reaction utilising the Shibasaki binol catalyst to obtain 70 as a major isomer in moderate yield.
O O OH
O
HO OTr
NC O OH
O OH
FmocHN O COOH
FmocHN 1. pyridine, DMAP, TrCl, DCM
2. DIBAL-H, DCM, -78 °C
1. Et3N, Ac2O, DMAP, DCM 2. TMSCN, BF3-Et2O, CH3CN
1. LiAlH4, Et2O 2. FmocOSu, DCM
Jones reagent, acetone
73 74 75
77 76
+ other isomer
Figure 21: furanoid amino acid by Chakraborty et al.
Another synthesis of a Fmoc protected furanoid δ-amino acid (figure 21) was proposed by Chakraborty in 2004.36 After protection of the primary hydroxyl group the starting material was reduced with DIBAL-H to the lactol 74. Acylation of the hydroxyl group was followed by treatment with trimethylsilyl cyanide in the presence of BF3 etherated to give 75 as a mixture of diastereomers. Reduction of the cyanide group followed by in situ N-Fmoc protection give the intermediate 76, which can be easily converted into the desired δ-amino acid by oxidation of the hydroxyl group with the Jones’ reagent.
Bn2N CHO R
Br Br
O O
n-BuLi, THF Bn2N
O O OH
R
BocHN OH
R 1. Pd(OH)2/C,
H2, MeOH 2. Boc2O,Et3N, MeOH
CSA, MeOH O
O
BocHN OH
R O OH
BocHN OH
R
1. TrisCl, Py, DCM 2. K2CO3, MeOH
BocHN O OMe
R
1. SO3-py, Et3N, DMSO, DCM 2. NaClO2, NaH2PO4 3. 2-methyl-2-butene, tBuOH 4. CH2N2, Et2O
78 79 80
82 81
83
OH
O
R = Me, CH2Ph, CH2Me, CH2OBn
Figure 22: route for C6-substituted furanoid amino acid by Chakraborty et al.
The same group in 200537 synthesised a large variety of C6-substituted furanoid amino acid with a completely different strategy (figure 22). Key step of the synthetic route is the selective sulfonylation of the primary hydroxyl group of 81 using 2,4,6-triisopropylbenzenesulfonyl chloride (TrisCl) which gave a sulfonate intermediate that was treated with anhydrous potassium carbonate to carry out an intramolecular ring closure reaction via an epoxide intermediate to give the tetrahydrofuran framework 82. The subsequent three steps oxidation of the primary hydroxyl group gives the desired δ-amino acid 83.
2.1.4 Six membered ring δδδδ-amino acids
Figure 23: route for pyranoid amino acid by Graf Von Roedern et Al.
In 1996 Graf von Roedern et al.38 synthesised a pyranoid δ-amino acid starting from glucosamine (figure 23). The anomeric hydroxyl was transformed to glycosyl bromide with acetyl bromide, which was treated with methanol and pyridine to give the β-methyl glycoside.
Benzyloxycarbonyl protection of the free amino group was followed by the methanolysis of acetyl group and selective oxidation of the primary hydroxyl group obtaining the desired δ-amino acid 88.
HO O HO
OH
NH3+Cl- OH
AcO O AcO
OAc
-Br+H3N
HO O HO
OH
NHCbz OMe
Br
AcO O AcO
OAc
NHCbz OMe
HO O HO
OH
NHCbz OMe O
AcBr
1. MeOH, pyridine 2.Cbz-Cl, DCM
MeOH, Me2EtN
Pt/C, O2
84 85
87 86
88
Figure 24: synthesis of pyranoid amino acids by Overkleeft et al.
In 1999, Overkleeft et al.39 synthesised a farnesyltransferase inhibitor based on sugar amino acids (figure 24). First step of the synthesis is a Ferrier rearrangement of 3,4,6-tri-O-acetyl-D-glucal 89 with trimethylsilyl cyanide and a catalityc amount of BF3 etherate to give a mixture of separable cyanides. Starting from the major isomer 90 after few steps was possible to obtain the 2 enantiomeric pure δ-amino acids 91 and 92.
AcO O AcO
AcO
AcO O AcO
CN
HO O
NHBoc HO O
CN PhthalHN HO
HO O
NHBoc HO O
HO O
COOH FmocHN
TMSCN, BF3-Et2O, DCM
1. H2, Pd/C, EtOAc, MeOH 2. Boc2O, Et3N, DMF 3. NaOMe, MeOH
TEMPO, KBr, NaOAc, NaOH 1. H2, Pd/C, EtOAc
2. NaOMe, MeOH 3. Phthalimide, DEAD, PPH3, THF
1. hydrazine, MeOH 2. Fmoc-Cl, DIPEA, DMF 3. HClconc, dioxane
89
91 90 93
92 94
O AcO AcO
OAc Br OAc
O HO HO
OH OH
NH2
O BnO BnO
OBn OMMtrt
NHBoc
O BnO BnO
OBn OH
NHBoc O
1. Hg(CN)2 2. LiAlH4, THF
1. Boc2O, THF, H2O
2. MMTr-Cl, TEA, DMAP, THF 3. Bn-Br, 18crown6, KOH, THF
1. 20% TFA/DCM, H2O 2. Fmoc-Cl, NaHCO3, THF 3. TEMPO, KBr, TBABr, NaOCl, DCM
95 95 97
98
Figure 25: Stockle’s pyranoid amino acid
A different synthetic route was proposed by Stockle et al.40. O-acetyl-α-glucopyranosyl bromide was reacted with Hg(CN)2 in melt to give the cyanide which was reduced to amine and then protected as tert-butoxy-carbonyl in situ with Boc anhydride. Carboxylic group was then introduced by a TEMPO oxidation of the primary hydroxyl group (figure 25).
BnO O
BnO O
OBn
OBn
BnO O
BnO N
OBn
OBn S O
BnO O
BnO N
H OBn
OBn S O BnO O
BnO NHFmoc
OBn
OBn
BnO O
BnO NHFmoc
OH
OBn O
R or S-butanesulfinimide, Ti(OiPr)4, DCM
MeMgBr, DCM
1. HCl, MeOH
2. FmocOSu, DIPEA, dioxane
1. ZnCl2, HOAc, Ac2O 2. HCl, MeOH
3. TEMPO, BAIB, DCM, H2O
99 100
102 101
103
Figure 26: sugar amino acid by Risseuw et al.
In 2007 Risseuw et al.41 proposed a synthesis of an alkylated sugar δ-amino acid by condensation of formil glucopyrasonide 99 with both R o S-butanesulfinimide in presence of Ti(OiPr)4, followed by the alkylation of the resulting imide with MeMgBr, giving the adduct 101 with an excellent diastereomeric excess (figure 26). Hydrolysis with methanolic HCl afforded the free amino group which was the protected by treatment with Fmoc succinimide. Selective debenzylation of the primary hydroxyl group followed by a TEMPO oxidation allowed to introduce the carboxylic function in the desired position.
2.1.5 Bicyclic δδδδ-amino acids
Between the bicyclic δ-amino acids, azabicyclo[X.Y.0]alkane amino acids or heteroatom
conformation which can surrogate a β-turn. Additionally, it is possible easily change some characteristics of the scaffold as rigidity or solubility by inclusion of substituents in different positions, insertion of heteroatoms or other modifications affecting the ring size. A quite general synthesis was proposed by Belvisi et al.42 in 2004, based on a radical approach (figure 27).
Starting from proline derivatives 104, after deprotection and alkylation of the amino group, was possible to obtain a small library of amides 105. Conversion of the alcohols to the corresponding brominated products 107 was performed by treatment with mesyl chloride followed by displacement with lithium bromide, whereas the selenides were prepared directly from the alcohols by treatment with N-phenylphtalimide and tributylphosphine.
N R1
COOtBu O NHAc
N R1
COOtBu O NHAc
Br n SePh
n
N COOtBu
HO n
R
N COOtBu
HO n
H
N R1
COOtBu O NHAc
HO n
104 105
106
107 108
H2, Pd/C
DCC, dehydroamino acid
1. MsCl, Et3N
2. LiBr, acetone N-PSP, Bu3P
n = 1, 2, 3 R = Bn, Cbz R1 = H, Me, Ph
N COOtBu O NHAc R1
N COOtBu O NHAc R1
n-Bu3SnH
N COOtBu O NHAc
N COOtBu O NHAc
N COOtBu O Me NHAc
Me
N COOtBu O NHAc
109 110 111
112 113 114
Figure 27: radical approach for azabicycloalkane by Belvisi et al.
